Re: pgsql: Add PSQL_WATCH_PAGER for psql's \watch command.

Thomas Munro <thomas(dot)munro(at)gmail(dot)com> writes:
> Here's a better patch.

I didn't like the way you'd done this as two independent tests;
the second one falsely reports success if there isn't actually
any declaration of sigwait. I see you hacked around that by
checking both results in c.h, but the printed result is still
very misleading. Plus it's a waste of time to run the second
test at all, if the first one fails.

Here's a revised patch that I've tested (albeit lightly) on
both HPUX and Solaris. I also got rid of the independent
check in thread_test.c, since that's always been a hack
that delivers a misleading complaint when it fails.

Personally, I'd skip adding the stanza in c.h in favor of just
making configure set USE_SIGWAIT directly, but perhaps you
see it differently.

regards, tom lane